MATLAB Benchmarks (Slurm)

Performance comparisons between the Hazel HPC's results and external results; Hazel meets or exceeds the expected performance.

Use parfor to Speed Up Monte-Carlo Code

External Links:
MATLAB Parallel Computing Benchmark
Hazel's performance:

• out.391753
• matlab_slurm_benchmark.m
• matlab_slurm_benchmark.sh
• The computational speedup increases with the number of workers. The speedup of Hazel is as good as or better than that of the external benchmark.






Hazel's Performance

External Benchmark's performance

Steps to reproduce this result:

Go to /share/$GROUP/$USER/

Create a directory and copy the matlab_slurm_benchmark.m and matlab_slurm_benchmark.sh files into it.

Then run the following command to run the job

sbatch matlab_slurm_benchmark.sh

The output file will be generated in the same folder once the job is completed.

Benchmarking A\b on the GPU

External Links:
MATLAB Parallel Computing Benchmark
Hazel's performance:

• out.345537
• matlab_gpu__slurm_benchmark.m
• matlab_gpu_slurm_benchmark.sh
• Compare the performance of the backslash operator on the CPU and the GPU for double precision. The performance of Hazel on A30 GPU is as good as or better than that of the external benchmark.






Hazel's Performance

External Benchmark's performance

Steps to reproduce this result:

Go to /share/$GROUP/$USER/

Create a directory and copy the matlab_gpu_slurm_benchmark.m and matlab_gpu_slurm_benchmark.sh files into it.

Then run the following command to run the job

sbatch matlab_gpu_slurm_benchmark.sh

The output file will be generated in the same folder once the job is completed.

Last modified: July 21 2026 13:22:59.